Compare commits

..

4 Commits

Author SHA1 Message Date
John Maguire 8589b76e1f Do not warn about local address failures on every update
localAddrs runs inside every SendUpdate, which fires on lighthouse.interval
and again on every network change via RebindUDPServer. Logging the failure
there meant a persistent failure warned every interval for the life of the
process, once per failing interface.

collectLocalAddrs now returns its failures instead of logging them, which
keeps it stateless and lets the tests assert on errors rather than log
output. The lighthouse holds the previous error and warns only when it
changes, demoting repeats to Debug, matching how handshake_manager handles
repeated send failures.
2026-07-24 17:53:55 -04:00
John Maguire 28cff022ee Link an android binary in build-test-mobile
The package builds only compile, so a linker-only failure could not fail
CI. anet relies on //go:linkname, which the linker rejects on Go 1.23+
without -checklinkname=0, so linking cmd/nebula for android both covers
that regression and records the flag requirement.
2026-07-24 17:53:55 -04:00
John Maguire 8cbee0e965 Advertise underlay addresses on Android
On Android 11+ the app sandbox denies bind() on netlink_route_socket, so
the stdlib's net.Interfaces fails with EACCES. localAddrs discarded that
error and returned an empty slice, so the node advertised no underlay
addresses and peers could only ever reach it at the address a lighthouse
observed. A device on the same LAN as a peer was unreachable at its LAN
address.

Split interface enumeration behind a build-tagged seam and use
github.com/wlynxg/anet on Android, which reads RTM_GETADDR from an
unbound socket. Interface addresses have to come from anet as well, since
net.Interface.Addrs goes back through the same denied path. Every other
platform keeps the net package implementation.

Stop discarding the enumeration errors, which are exceptional now that
the sandbox case is handled.

anet needs -ldflags=-checklinkname=0 on Go 1.23+. Nebula ships no Android
binaries, so build-test-mobile is unaffected, but consumers linking
Android artifacts will need the flag.
2026-07-24 15:06:29 -04:00
Nate Brown 72bf111209 Add an e2e Drop exit type and a roaming recovery measurement (#1819)
smoke-extra / freebsd-amd64 (push) Failing after 15s
smoke-extra / linux-amd64-ipv6disable (push) Failing after 15s
smoke-extra / netbsd-amd64 (push) Failing after 14s
smoke-extra / openbsd-amd64 (push) Failing after 15s
smoke-extra / linux-386 (push) Failing after 16s
smoke / Run multi node smoke test (push) Failing after 1m37s
Build and test / Static checks (push) Successful in 18s
Build and test / Test linux (push) Failing after 58s
Build and test / Test linux-boringcrypto (push) Failing after 2m45s
Build and test / Test linux-pkcs11 (push) Failing after 2m10s
Build and test / Cross-build linux-arm (push) Successful in 3m11s
Build and test / Cross-build linux-mips (push) Successful in 3m53s
Build and test / Cross-build linux-other (push) Successful in 3m16s
Build and test / Cross-build windows (push) Successful in 1m2s
Build and test / Cross-build freebsd (push) Successful in 1m36s
Build and test / Cross-build netbsd (push) Successful in 1m36s
Build and test / Cross-build openbsd (push) Successful in 1m37s
Build and test / Cross-build mobile (push) Successful in 3m23s
smoke-extra / Run windows smoke test (push) Has been cancelled
Build and test / Test macos (push) Has been cancelled
Build and test / Test windows (push) Has been cancelled
Build and test / CI status (push) Has been cancelled
2026-07-23 17:02:02 -05:00
31 changed files with 400 additions and 718 deletions
-8
View File
@@ -60,12 +60,4 @@ jobs:
working-directory: ./.github/workflows/smoke
run: NAME="smoke-p256" ./smoke.sh
- name: setup docker image for multiport
working-directory: ./.github/workflows/smoke
run: NAME="smoke-multiport" MULTIPORT_TX=true MULTIPORT_RX=true MULTIPORT_HANDSHAKE=true ./build.sh
- name: run smoke
working-directory: ./.github/workflows/smoke
run: NAME="smoke-multiport" ./smoke.sh
timeout-minutes: 10
-4
View File
@@ -48,10 +48,6 @@ listen:
tun:
dev: ${TUN_DEV:-tun0}
multiport:
tx_enabled: ${MULTIPORT_TX:-false}
rx_enabled: ${MULTIPORT_RX:-false}
tx_handshake: ${MULTIPORT_HANDSHAKE:-false}
firewall:
inbound_action: reject
+3 -4
View File
@@ -222,11 +222,14 @@ test-cov-html:
go test -coverprofile=coverage.out
go tool cover -html=coverage.out
# The package builds only compile. The final line links an android binary so a linker-only failure,
# such as the //go:linkname reference anet makes, cannot pass CI.
build-test-mobile:
GOARCH=amd64 GOOS=ios go build $(shell go list ./... | grep -v '/cmd/\|/examples/')
GOARCH=arm64 GOOS=ios go build $(shell go list ./... | grep -v '/cmd/\|/examples/')
GOARCH=amd64 GOOS=android go build $(shell go list ./... | grep -v '/cmd/\|/examples/')
GOARCH=arm64 GOOS=android go build $(shell go list ./... | grep -v '/cmd/\|/examples/')
GOARCH=arm64 GOOS=android go build -ldflags=-checklinkname=0 -o /dev/null ${NEBULA_CMD_PATH}
bench:
go test -bench=.
@@ -268,10 +271,6 @@ smoke-relay-docker: bin-docker
cd .github/workflows/smoke/ && ./build-relay.sh
cd .github/workflows/smoke/ && ./smoke-relay.sh
smoke-multiport-docker: bin-docker
cd .github/workflows/smoke/ && NAME="smoke-multiport" MULTIPORT_TX=true MULTIPORT_RX=true MULTIPORT_HANDSHAKE=true ./build.sh
cd .github/workflows/smoke/ && NAME="smoke-multiport" ./smoke.sh
smoke-docker-ipv6: export SMOKE_OVERLAY_IPV6 = 1
smoke-docker-ipv6: smoke-docker
-10
View File
@@ -1,10 +0,0 @@
package config
type MultiPortConfig struct {
Tx bool
Rx bool
TxBasePort uint16
TxPorts int
TxHandshake bool
TxHandshakeDelay int64
}
-3
View File
@@ -8,7 +8,6 @@ import (
"github.com/flynn/noise"
"github.com/slackhq/nebula/cert"
ct "github.com/slackhq/nebula/cert_test"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/handshake"
"github.com/slackhq/nebula/header"
"github.com/stretchr/testify/assert"
@@ -56,7 +55,6 @@ func runTestHandshake(t *testing.T) (initR, respR *handshake.Result) {
cert.Version2, initCreds, verifier,
func() (uint32, error) { return 1000, nil },
true, header.HandshakeIXPSK0,
config.MultiPortConfig{},
)
require.NoError(t, err)
@@ -64,7 +62,6 @@ func runTestHandshake(t *testing.T) (initR, respR *handshake.Result) {
cert.Version2, respCreds, verifier,
func() (uint32, error) { return 2000, nil },
false, header.HandshakeIXPSK0,
config.MultiPortConfig{},
)
require.NoError(t, err)
+136
View File
@@ -0,0 +1,136 @@
//go:build e2e_testing
// +build e2e_testing
package e2e
import (
"testing"
"time"
"github.com/slackhq/nebula"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/cert_test"
"github.com/slackhq/nebula/e2e/router"
"github.com/slackhq/nebula/udp"
)
// TestRecoveryTiming measures how long a tunnel takes to come back after the peer stops accepting our traffic,
// which is what a laptop waking on a new network looks like from the peer's side: its NAT has no state for where
// we are now, so everything we send disappears.
//
// It is a measurement, not a pass/fail assertion. Recovery is timed to the moment the peer punches back at us,
// since that is when its NAT opens and the tunnel is usable again.
//
// go test -tags e2e_testing -v -run TestRecoveryTiming ./e2e/
func TestRecoveryTiming(t *testing.T) {
for _, tc := range []struct {
name string
rebind bool
}{
{"no trigger", false},
{"rebind counter", true},
} {
t.Run(tc.name, func(t *testing.T) {
d, lost := measureRecovery(t, tc.rebind)
t.Logf("RESULT %-16s recovered in %-9v (%d packets lost)", tc.name, d.Round(time.Millisecond), lost)
})
}
}
// measureRecovery returns how long until the peer punched back, and how many of our packets died meanwhile. When
// rebind is true we call RebindUDPServer once the tunnel goes dark, which is what the darwin network change
// monitor does and what iOS has always done. When false, nothing tells nebula anything is wrong.
func measureRecovery(t *testing.T, rebind bool) (time.Duration, int) {
t.Helper()
ca, _, caKey, _ := cert_test.NewTestCaCert(cert.Version2, cert.Curve_CURVE25519, time.Now(), time.Now().Add(10*time.Minute), nil, nil, []string{})
lhControl, lhVpnIpNet, lhUdpAddr, _ := newSimpleServer(cert.Version2, ca, caKey, "lh", "10.128.0.1/24", m{
"lighthouse": m{"am_lighthouse": true},
})
peerCfg := m{
"lighthouse": m{
"hosts": []any{lhVpnIpNet[0].Addr().String()},
"interval": 600,
"local_allow_list": m{
"10.0.0.0/24": true,
"::/0": false,
},
},
"static_host_map": m{
lhVpnIpNet[0].Addr().String(): []any{lhUdpAddr.String()},
},
}
myControl, myVpnIpNet, myUdpAddr, _ := newSimpleServer(cert.Version2, ca, caKey, "me", "10.128.0.2/24", peerCfg)
theirControl, theirVpnIpNet, theirUdpAddr, _ := newSimpleServer(cert.Version2, ca, caKey, "them", "10.128.0.3/24", peerCfg)
r := router.NewR(t, lhControl, myControl, theirControl)
defer r.RenderFlow()
defer func() {
lhControl.Stop()
myControl.Stop()
theirControl.Stop()
}()
lhControl.Start()
myControl.Start()
theirControl.Start()
r.RouteFor(time.Millisecond * 500)
myControl.InjectLightHouseAddr(theirVpnIpNet[0].Addr(), theirUdpAddr)
theirControl.InjectLightHouseAddr(myVpnIpNet[0].Addr(), myUdpAddr)
myControl.InjectTunPacket(BuildTunUDPPacket(theirVpnIpNet[0].Addr(), 80, myVpnIpNet[0].Addr(), 80, []byte("establish")))
r.RouteFor(time.Second)
if myControl.GetHostInfoByVpnAddr(theirVpnIpNet[0].Addr(), false) == nil {
t.Fatal("failed to establish the tunnel we are measuring")
}
r.RouteFor(time.Millisecond * 500)
// From here the peer's NAT has no state for us, everything we send it disappears
start := time.Now()
blackholed := 0
var recovered time.Duration
if rebind {
myControl.RebindUDPServer()
}
// Keep the tun busy the way someone retrying a stalled connection would
stop := make(chan struct{})
defer close(stop)
go func() {
tick := time.NewTicker(time.Millisecond * 200)
defer tick.Stop()
for {
select {
case <-stop:
return
case <-tick.C:
myControl.InjectTunPacket(BuildTunUDPPacket(
theirVpnIpNet[0].Addr(), 80, myVpnIpNet[0].Addr(), 80, []byte("retry")))
}
}
}()
r.RouteForAllExitFuncOrTimeout(time.Second*30, func(p *udp.Packet, c *nebula.Control) router.ExitType {
if c == theirControl && p.From == myControl.GetUDPAddr() {
blackholed++
return router.Drop
}
// The peer reaching us directly is the moment its NAT opened, whether that is a punch or a handshake
if c == myControl && p.From == theirUdpAddr {
recovered = time.Since(start)
return router.RouteAndExit
}
return router.KeepRouting
})
if recovered == 0 {
t.Fatalf("no recovery within 30s (%d packets blackholed)", blackholed)
}
return recovered, blackholed
}
+19 -2
View File
@@ -153,6 +153,9 @@ const (
ExitNow ExitType = 1
// RouteAndExit routes this packet and exits immediately afterwards
RouteAndExit ExitType = 2
// Drop discards this packet without delivering it and keeps routing. Use it to simulate a blackhole, such as
// a restrictive NAT refusing traffic from an address it has not seen.
Drop ExitType = 3
)
type ExitFunc func(packet *udp.Packet, receiver *nebula.Control) ExitType
@@ -163,7 +166,9 @@ type ExitFunc func(packet *udp.Packet, receiver *nebula.Control) ExitType
func NewR(t testing.TB, controls ...*nebula.Control) *R {
ctx, cancel := context.WithCancel(context.Background())
if err := os.MkdirAll("mermaid", 0755); err != nil {
// t.Name() contains a slash for subtests, so the flow log can land in a nested directory
fn := filepath.Join("mermaid", fmt.Sprintf("%s.md", t.Name()))
if err := os.MkdirAll(filepath.Dir(fn), 0755); err != nil {
panic(err)
}
@@ -174,7 +179,7 @@ func NewR(t testing.TB, controls ...*nebula.Control) *R {
outNat: make(map[outNatKey]netip.AddrPort),
flow: []flowEntry{},
ignoreFlows: []ignoreFlow{},
fn: filepath.Join("mermaid", fmt.Sprintf("%s.md", t.Name())),
fn: fn,
t: t,
cancelRender: cancel,
}
@@ -687,6 +692,10 @@ func (r *R) RouteExitFunc(sender *nebula.Control, whatDo ExitFunc) {
p.Release()
return
case Drop:
// Record it so the flow log shows the attempt, but never hand it to the receiver
r.unlockedInjectFlow(sender, receiver, p, false)
case KeepRouting:
fp := r.unlockedInjectFlow(sender, receiver, p, false)
receiver.InjectUDPPacket(p)
@@ -779,6 +788,10 @@ func (r *R) RouteForAllExitFuncOrTimeout(timeout time.Duration, whatDo ExitFunc)
p.Release()
return true
case Drop:
// Record it so the flow log shows the attempt, but never hand it to the receiver
r.unlockedInjectFlow(cm[x], receiver, p, false)
case KeepRouting:
fp := r.unlockedInjectFlow(cm[x], receiver, p, false)
receiver.InjectUDPPacket(p)
@@ -884,6 +897,10 @@ func (r *R) RouteForAllExitFunc(whatDo ExitFunc) {
p.Release()
return
case Drop:
// Record it so the flow log shows the attempt, but never hand it to the receiver
r.unlockedInjectFlow(cm[x], receiver, p, false)
case KeepRouting:
fp := r.unlockedInjectFlow(cm[x], receiver, p, false)
receiver.InjectUDPPacket(p)
-41
View File
@@ -328,47 +328,6 @@ tun:
# SO_RCVBUFFORCE is used to avoid having to raise the system wide max
#use_system_route_table_buffer_size: 0
# EXPERIMENTAL: This option may change or disappear in the future.
# Multiport spreads outgoing UDP packets across multiple UDP send ports,
# which allows nebula to work around any issues on the underlay network.
# Some example issues this could work around:
# - UDP rate limits on a per flow basis.
# - Partial underlay network failure in which some flows work and some don't
# Agreement is done during the handshake to decide if multiport mode will
# be used for a given tunnel (one side must have tx_enabled set, the other
# side must have rx_enabled set)
#
# NOTE: you cannot use multiport on a host if you are relying on UDP hole
# punching to get through a NAT or firewall.
#
# NOTE: Linux only (uses raw sockets to send). Also currently only works
# with IPv4 underlay network remotes.
#
# The default values are listed below:
#multiport:
# This host support sending via multiple UDP ports.
#tx_enabled: false
#
# This host supports receiving packets sent from multiple UDP ports.
#rx_enabled: false
#
# How many UDP ports to use when sending. The lowest source port will be
# listen.port and go up to (but not including) listen.port + tx_ports.
#tx_ports: 100
#
# NOTE: All of your hosts must be running a version of Nebula that supports
# multiport if you want to enable this feature. Older versions of Nebula
# will be confused by these multiport handshakes.
#
# If handshakes are not getting a response, attempt to transmit handshakes
# using random UDP source ports (to get around partial underlay network
# failures).
#tx_handshake: false
#
# How many unresponded handshakes we should send before we attempt to
# send multiport handshakes.
#tx_handshake_delay: 2
# Configure logging level
logging:
# trace, debug, info, warn, or error. Default is info and is reloadable.
-28
View File
@@ -3,7 +3,6 @@ package firewall
import (
"encoding/json"
"fmt"
mathrand "math/rand"
"net/netip"
)
@@ -66,30 +65,3 @@ func (fp Packet) MarshalJSON() ([]byte, error) {
"Fragment": fp.Fragment,
})
}
// UDPSendPort calculates the UDP port to send from when using multiport mode.
// The result will be from [0, numBuckets)
func (fp Packet) UDPSendPort(numBuckets int) uint16 {
if numBuckets <= 1 {
return 0
}
// If there is no port (like an ICMP packet), pick a random UDP send port
if fp.LocalPort == 0 {
return uint16(mathrand.Intn(numBuckets))
}
// A decent enough 32bit hash function
// Prospecting for Hash Functions
// - https://nullprogram.com/blog/2018/07/31/
// - https://github.com/skeeto/hash-prospector
// [16 21f0aaad 15 d35a2d97 15] = 0.10760229515479501
x := (uint32(fp.LocalPort) << 16) | uint32(fp.RemotePort)
x ^= x >> 16
x *= 0x21f0aaad
x ^= x >> 15
x *= 0xd35a2d97
x ^= x >> 15
return uint16(x) % uint16(numBuckets)
}
+1
View File
@@ -22,6 +22,7 @@ require (
github.com/stefanberger/go-pkcs11uri v0.0.0-20230803200340-78284954bff6
github.com/stretchr/testify v1.11.1
github.com/vishvananda/netlink v1.3.1
github.com/wlynxg/anet v0.0.5
go.uber.org/goleak v1.3.0
go.yaml.in/yaml/v3 v3.0.4
golang.org/x/crypto v0.54.0
+2
View File
@@ -149,6 +149,8 @@ github.com/vishvananda/netlink v1.3.1 h1:3AEMt62VKqz90r0tmNhog0r/PpWKmrEShJU0wJW
github.com/vishvananda/netlink v1.3.1/go.mod h1:ARtKouGSTGchR8aMwmkzC0qiNPrrWO5JS/XMVl45+b4=
github.com/vishvananda/netns v0.0.5 h1:DfiHV+j8bA32MFM7bfEunvT8IAqQ/NzSJHtcmW5zdEY=
github.com/vishvananda/netns v0.0.5/go.mod h1:SpkAiCQRtJ6TvvxPnOSyH3BMl6unz3xZlaprSwhNNJM=
github.com/wlynxg/anet v0.0.5 h1:J3VJGi1gvo0JwZ/P1/Yc/8p63SoW98B5dHkYDmpgvvU=
github.com/wlynxg/anet v0.0.5/go.mod h1:eay5PRQr7fIVAMbTbchTnO9gG65Hg/uYGdc7mguHxoA=
github.com/yuin/goldmark v1.1.27/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
github.com/yuin/goldmark v1.2.1/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
go.uber.org/goleak v1.3.0 h1:2K3zAYmnTNqV73imy9J1T3WC+gmCePx2hEGkimedGto=
+2 -10
View File
@@ -24,14 +24,6 @@ message NebulaHandshakeDetails {
uint64 Cookie = 4 [deprecated = true];
uint64 Time = 5;
uint32 CertVersion = 8;
MultiPortDetails InitiatorMultiPort = 6;
MultiPortDetails ResponderMultiPort = 7;
}
message MultiPortDetails {
bool RxSupported = 1;
bool TxSupported = 2;
uint32 BasePort = 3;
uint32 TotalPorts = 4;
// reserved for WIP multiport
reserved 6, 7;
}
-2
View File
@@ -8,7 +8,6 @@ import (
"github.com/flynn/noise"
"github.com/slackhq/nebula/cert"
ct "github.com/slackhq/nebula/cert_test"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/header"
"github.com/stretchr/testify/require"
)
@@ -72,7 +71,6 @@ func newTestMachine(
cs.version, cs.getCredential,
verifier, func() (uint32, error) { return localIndex, nil },
initiator, header.HandshakeIXPSK0,
config.MultiPortConfig{},
)
require.NoError(t, err)
return m
+1 -43
View File
@@ -3,13 +3,11 @@ package handshake
import (
"bytes"
"fmt"
"math"
"slices"
"time"
"github.com/flynn/noise"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/header"
)
@@ -41,10 +39,6 @@ type Result struct {
HandshakeTime uint64
MessageIndex uint64 // number of messages exchanged during the handshake
Initiator bool
MultiportRx bool
MultiportTx bool
MultiportBasePort uint16
}
// Machine drives a Noise handshake through N messages. It handles Noise
@@ -73,8 +67,6 @@ type Machine struct {
remoteCertSet bool
payloadSet bool
failed bool
multiport config.MultiPortConfig
}
// NewMachine creates a handshake state machine. The subtype determines both
@@ -88,7 +80,6 @@ func NewMachine(
allocIndex IndexAllocator,
initiator bool,
subtype header.MessageSubType,
multiport config.MultiPortConfig,
) (*Machine, error) {
info, err := subtypeInfoFor(subtype)
if err != nil {
@@ -117,8 +108,6 @@ func NewMachine(
Initiator: initiator,
Cipher: cred.cipherSuite,
},
multiport: multiport,
}, nil
}
@@ -309,7 +298,7 @@ func (m *Machine) processPayload(msg []byte, flags msgFlags) error {
}
// Assert the payload contains exactly what we expect
hasPayloadData := payload.InitiatorIndex != 0 || payload.ResponderIndex != 0 || payload.Time != 0 || payload.InitiatorMultiPort != nil || payload.ResponderMultiPort != nil
hasPayloadData := payload.InitiatorIndex != 0 || payload.ResponderIndex != 0 || payload.Time != 0
if hasPayloadData != flags.expectsPayload {
m.failed = true
return ErrUnexpectedContent
@@ -326,22 +315,8 @@ func (m *Machine) processPayload(msg []byte, flags msgFlags) error {
var remoteIndex uint32
if m.result.Initiator {
remoteIndex = payload.ResponderIndex
if payload.ResponderMultiPort != nil {
m.result.MultiportRx = payload.ResponderMultiPort.RxSupported
m.result.MultiportTx = payload.ResponderMultiPort.TxSupported
if payload.ResponderMultiPort.BasePort <= math.MaxUint16 {
m.result.MultiportBasePort = uint16(payload.ResponderMultiPort.BasePort)
}
}
} else {
remoteIndex = payload.InitiatorIndex
if payload.InitiatorMultiPort != nil {
m.result.MultiportRx = payload.InitiatorMultiPort.RxSupported
m.result.MultiportTx = payload.InitiatorMultiPort.TxSupported
if payload.InitiatorMultiPort.BasePort <= math.MaxUint16 {
m.result.MultiportBasePort = uint16(payload.InitiatorMultiPort.BasePort)
}
}
}
// The payload presence check above can be satisfied by Time alone, so a payload
// could still carry a zero index here. We need to reject it.
@@ -422,28 +397,11 @@ func (m *Machine) marshalOutgoing(flags msgFlags) ([]byte, error) {
if m.result.Initiator {
p.InitiatorIndex = m.result.LocalIndex
if m.multiport.Rx || m.multiport.Tx {
p.InitiatorMultiPort = &PayloadMultiPortDetails{
RxSupported: m.multiport.Rx,
TxSupported: m.multiport.Tx,
BasePort: uint32(m.multiport.TxBasePort),
TotalPorts: uint32(m.multiport.TxPorts),
}
}
} else {
p.ResponderIndex = m.result.LocalIndex
p.InitiatorIndex = m.result.RemoteIndex
if m.multiport.Rx || m.multiport.Tx {
p.ResponderMultiPort = &PayloadMultiPortDetails{
RxSupported: m.multiport.Rx,
TxSupported: m.multiport.Tx,
BasePort: uint32(m.multiport.TxBasePort),
TotalPorts: uint32(m.multiport.TxPorts),
}
}
}
p.Time = uint64(time.Now().UnixNano())
}
if flags.expectsCert {
cred := m.getCred(m.myVersion)
-3
View File
@@ -8,7 +8,6 @@ import (
"github.com/flynn/noise"
"github.com/slackhq/nebula/cert"
ct "github.com/slackhq/nebula/cert_test"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/noiseutil"
"github.com/stretchr/testify/assert"
@@ -445,7 +444,6 @@ func TestMachineThreeMessagePattern(t *testing.T) {
initCS.getCredential, v,
func() (uint32, error) { return 1000, nil },
true, header.HandshakeXXPSK0,
config.MultiPortConfig{},
)
require.NoError(t, err)
@@ -454,7 +452,6 @@ func TestMachineThreeMessagePattern(t *testing.T) {
respCS.getCredential, v,
func() (uint32, error) { return 2000, nil },
false, header.HandshakeXXPSK0,
config.MultiPortConfig{},
)
require.NoError(t, err)
-139
View File
@@ -20,16 +20,6 @@ type Payload struct {
ResponderIndex uint32
Time uint64
CertVersion uint32
InitiatorMultiPort *PayloadMultiPortDetails
ResponderMultiPort *PayloadMultiPortDetails
}
type PayloadMultiPortDetails struct {
RxSupported bool
TxSupported bool
BasePort uint32
TotalPorts uint32
}
// Proto field numbers for NebulaHandshakeDetails
@@ -39,17 +29,6 @@ const (
fieldResponderIndex = 3 // uint32
fieldTime = 5 // uint64
fieldCertVersion = 8 // uint32
fieldInitiatorMultiPort = 6 // MultiPortDetails
fieldResponderMultiPort = 7 // MultiPortDetails
)
// Proto field numbers for MultiPortDetails
const (
fieldMultiportRxSupported = 1 // bool
fieldMultiportTxSupported = 2 // bool
fieldMultiportBasePort = 3 // uint32
fieldMultiportTotalPorts = 4 // uint32
)
// MarshalPayload encodes a handshake payload in protobuf wire format compatible
@@ -78,16 +57,6 @@ func MarshalPayload(out []byte, p Payload) []byte {
details = protowire.AppendTag(details, fieldCertVersion, protowire.VarintType)
details = protowire.AppendVarint(details, uint64(p.CertVersion))
}
if p.InitiatorMultiPort != nil {
details = protowire.AppendTag(details, fieldInitiatorMultiPort, protowire.BytesType)
details = protowire.AppendVarint(details, uint64(p.InitiatorMultiPort.size()))
details = p.InitiatorMultiPort.marshal(details)
}
if p.ResponderMultiPort != nil {
details = protowire.AppendTag(details, fieldResponderMultiPort, protowire.BytesType)
details = protowire.AppendVarint(details, uint64(p.ResponderMultiPort.size()))
details = p.ResponderMultiPort.marshal(details)
}
out = protowire.AppendTag(out, 1, protowire.BytesType)
out = protowire.AppendBytes(out, details)
@@ -95,23 +64,6 @@ func MarshalPayload(out []byte, p Payload) []byte {
return out
}
func (p PayloadMultiPortDetails) marshal(details []byte) []byte {
details = protowire.AppendTag(details, fieldMultiportRxSupported, protowire.VarintType)
details = protowire.AppendVarint(details, protowire.EncodeBool(p.RxSupported))
details = protowire.AppendTag(details, fieldMultiportTxSupported, protowire.VarintType)
details = protowire.AppendVarint(details, protowire.EncodeBool(p.TxSupported))
details = protowire.AppendTag(details, fieldMultiportBasePort, protowire.VarintType)
details = protowire.AppendVarint(details, uint64(p.BasePort))
details = protowire.AppendTag(details, fieldMultiportTotalPorts, protowire.VarintType)
details = protowire.AppendVarint(details, uint64(p.TotalPorts))
return details
}
func (p PayloadMultiPortDetails) size() int {
return 4 + 2 + protowire.SizeVarint(uint64(p.BasePort)) + protowire.SizeVarint(uint64(p.TotalPorts))
}
// UnmarshalPayload decodes a protobuf-encoded NebulaHandshake message.
func UnmarshalPayload(b []byte) (Payload, error) {
var p Payload
@@ -209,97 +161,6 @@ func unmarshalPayloadDetails(p *Payload, b []byte) error {
}
p.CertVersion = uint32(v)
b = b[n:]
case fieldInitiatorMultiPort:
if typ != protowire.BytesType {
return errInvalidHandshakeDetails
}
d, n := protowire.ConsumeBytes(b)
if n < 0 {
return errInvalidHandshakeMessage
}
b = b[n:]
p.InitiatorMultiPort = new(PayloadMultiPortDetails)
if err := unmarshalPayloadMultiPortDetails(p.InitiatorMultiPort, d); err != nil {
return err
}
case fieldResponderMultiPort:
if typ != protowire.BytesType {
return errInvalidHandshakeDetails
}
d, n := protowire.ConsumeBytes(b)
if n < 0 {
return errInvalidHandshakeMessage
}
b = b[n:]
p.ResponderMultiPort = new(PayloadMultiPortDetails)
if err := unmarshalPayloadMultiPortDetails(p.ResponderMultiPort, d); err != nil {
return err
}
default:
n := protowire.ConsumeFieldValue(num, typ, b)
if n < 0 {
return errInvalidHandshakeDetails
}
b = b[n:]
}
}
return nil
}
func unmarshalPayloadMultiPortDetails(p *PayloadMultiPortDetails, b []byte) error {
for len(b) > 0 {
num, typ, n := protowire.ConsumeTag(b)
if n < 0 {
return errInvalidHandshakeDetails
}
b = b[n:]
// For known field numbers, reject any non-matching wire type as a
// hard error rather than silently skipping. The caller will catch
// missing-field cases downstream, but a wire-type mismatch on a tag
// we know is a peer protocol violation worth flagging here.
// Repeated occurrences of a singular field follow proto3 last-wins.
switch num {
case fieldMultiportRxSupported:
if typ != protowire.VarintType {
return errInvalidHandshakeDetails
}
v, n := protowire.ConsumeVarint(b)
if n < 0 || v > math.MaxUint32 {
return errInvalidHandshakeDetails
}
p.RxSupported = protowire.DecodeBool(v)
b = b[n:]
case fieldMultiportTxSupported:
if typ != protowire.VarintType {
return errInvalidHandshakeDetails
}
v, n := protowire.ConsumeVarint(b)
if n < 0 || v > math.MaxUint32 {
return errInvalidHandshakeDetails
}
p.TxSupported = protowire.DecodeBool(v)
b = b[n:]
case fieldMultiportBasePort:
if typ != protowire.VarintType {
return errInvalidHandshakeDetails
}
v, n := protowire.ConsumeVarint(b)
if n < 0 || v > math.MaxUint32 {
return errInvalidHandshakeDetails
}
p.BasePort = uint32(v)
b = b[n:]
case fieldMultiportTotalPorts:
if typ != protowire.VarintType {
return errInvalidHandshakeDetails
}
v, n := protowire.ConsumeVarint(b)
if n < 0 || v > math.MaxUint32 {
return errInvalidHandshakeDetails
}
p.TotalPorts = uint32(v)
b = b[n:]
default:
n := protowire.ConsumeFieldValue(num, typ, b)
if n < 0 {
+16 -16
View File
@@ -117,24 +117,24 @@ func TestPayloadUnknownFields(t *testing.T) {
assert.Equal(t, uint32(88), got.ResponderIndex)
})
// t.Run("reserved fields 6 and 7 are skipped", func(t *testing.T) {
// // Fields 6 and 7 are reserved in the proto definition
// var details []byte
// details = protowire.AppendTag(details, fieldInitiatorIndex, protowire.VarintType)
// details = protowire.AppendVarint(details, 100)
// details = protowire.AppendTag(details, 6, protowire.VarintType)
// details = protowire.AppendVarint(details, 1)
// details = protowire.AppendTag(details, 7, protowire.VarintType)
// details = protowire.AppendVarint(details, 2)
t.Run("reserved fields 6 and 7 are skipped", func(t *testing.T) {
// Fields 6 and 7 are reserved in the proto definition
var details []byte
details = protowire.AppendTag(details, fieldInitiatorIndex, protowire.VarintType)
details = protowire.AppendVarint(details, 100)
details = protowire.AppendTag(details, 6, protowire.VarintType)
details = protowire.AppendVarint(details, 1)
details = protowire.AppendTag(details, 7, protowire.VarintType)
details = protowire.AppendVarint(details, 2)
// var data []byte
// data = protowire.AppendTag(data, 1, protowire.BytesType)
// data = protowire.AppendBytes(data, details)
var data []byte
data = protowire.AppendTag(data, 1, protowire.BytesType)
data = protowire.AppendBytes(data, details)
// got, err := UnmarshalPayload(data)
// require.NoError(t, err)
// assert.Equal(t, uint32(100), got.InitiatorIndex)
// })
got, err := UnmarshalPayload(data)
require.NoError(t, err)
assert.Equal(t, uint32(100), got.InitiatorIndex)
})
}
func TestPayloadBytesConsumed(t *testing.T) {
-55
View File
@@ -14,7 +14,6 @@ import (
"github.com/rcrowley/go-metrics"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/handshake"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/udp"
@@ -72,9 +71,6 @@ type HandshakeManager struct {
f *Interface
l *slog.Logger
multiPort config.MultiPortConfig
udpRaw *udp.RawConn
// can be used to trigger outbound handshake for the given vpnIp
trigger chan netip.Addr
}
@@ -295,7 +291,6 @@ func (hm *HandshakeManager) handleOutbound(vpnIp netip.Addr, lighthouseTriggered
// Send the handshake to all known ips, stage 2 takes care of assigning the hostinfo.remote based on the first to reply
var sentTo []netip.AddrPort
var sentMultiport bool
hostinfo.remotes.ForEach(hm.mainHostMap.GetPreferredRanges(), func(addr netip.AddrPort, _ bool) {
hm.messageMetrics.Tx(header.Handshake, hh.machine.Subtype(), 1)
err := hm.outside.WriteTo(stage0, addr)
@@ -316,29 +311,6 @@ func (hm *HandshakeManager) handleOutbound(vpnIp netip.Addr, lighthouseTriggered
} else {
sentTo = append(sentTo, addr)
}
// Attempt a multiport handshake if we are past the TxHandshakeDelay attempts
if hm.multiPort.TxHandshake && hm.udpRaw != nil && hh.counter >= hm.multiPort.TxHandshakeDelay {
sentMultiport = true
// We need to re-allocate with 8 bytes at the start of SOCK_RAW
raw := hostinfo.HandshakePacket[0x80]
if raw == nil {
raw = make([]byte, len(hostinfo.HandshakePacket[0])+udp.RawOverhead)
copy(raw[udp.RawOverhead:], hostinfo.HandshakePacket[0])
hostinfo.HandshakePacket[0x80] = raw
}
hm.messageMetrics.Tx(header.Handshake, header.MessageSubType(hostinfo.HandshakePacket[0][1]), 1)
err = hm.udpRaw.WriteTo(raw, udp.RandomSendPort.UDPSendPort(hm.multiPort.TxPorts), addr)
if err != nil {
hostinfo.logger(hm.l).Error("Failed to send handshake message",
"error", err,
"udpAddr", addr,
"initiatorIndex", hostinfo.localIndexId,
"handshake", hsFields,
)
}
}
})
// Don't be too noisy or confusing if we fail to send a handshake - if we don't get through we'll eventually log a timeout,
@@ -348,7 +320,6 @@ func (hm *HandshakeManager) handleOutbound(vpnIp netip.Addr, lighthouseTriggered
"udpAddrs", sentTo,
"initiatorIndex", hostinfo.localIndexId,
"handshake", hsFields,
"multiportHandshake", sentMultiport,
)
} else if hm.l.Enabled(context.Background(), slog.LevelDebug) {
hostinfo.logger(hm.l).Debug("Handshake message sent",
@@ -701,7 +672,6 @@ func (hm *HandshakeManager) buildStage0Packet(hh *HandshakeHostInfo) bool {
v, cs.GetCredential,
hm.certVerifier(), func() (uint32, error) { return hm.allocateIndex(hh) },
true, header.HandshakeIXPSK0,
hm.multiPort,
)
if err != nil {
hm.f.l.Error("Failed to create handshake machine",
@@ -743,7 +713,6 @@ func (hm *HandshakeManager) beginHandshake(via ViaSender, packet []byte, h *head
v, cs.GetCredential,
hm.certVerifier(), func() (uint32, error) { return generateIndex(f.l) },
false, header.HandshakeIXPSK0,
hm.multiPort,
)
if err != nil {
f.l.Error("Failed to create handshake machine", "from", via, "error", err)
@@ -768,12 +737,6 @@ func (hm *HandshakeManager) beginHandshake(via ViaSender, packet []byte, h *head
return
}
if !via.IsRelayed && result.MultiportTx && result.MultiportBasePort != via.UdpAddr.Port() {
// The other side sent us a handshake from a different port, make sure
// we send responses back to the BasePort
via.UdpAddr = netip.AddrPortFrom(via.UdpAddr.Addr(), result.MultiportBasePort)
}
remoteCert := result.RemoteCert
if remoteCert == nil {
f.l.Error("Handshake did not produce a peer certificate", "from", via)
@@ -798,8 +761,6 @@ func (hm *HandshakeManager) beginHandshake(via ViaSender, packet []byte, h *head
relayForByAddr: map[netip.Addr]*Relay{},
relayForByIdx: map[uint32]*Relay{},
},
multiportTx: hm.multiPort.Tx && result.MultiportRx,
multiportRx: hm.multiPort.Rx && result.MultiportTx,
}
msg := "Handshake message received"
@@ -816,8 +777,6 @@ func (hm *HandshakeManager) beginHandshake(via ViaSender, packet []byte, h *head
"initiatorIndex", result.RemoteIndex,
"responderIndex", result.LocalIndex,
"handshake", m{"stage": uint64(machine.MessageIndex()), "style": header.SubTypeName(header.Handshake, machine.Subtype())},
"multiportTx", hostinfo.multiportTx,
"multiportRx", hostinfo.multiportRx,
)
// packet aliases the listener's incoming buffer, so this copy must stay.
@@ -908,14 +867,6 @@ func (hm *HandshakeManager) continueHandshake(via ViaSender, hh *HandshakeHostIn
return
}
if !via.IsRelayed && result.MultiportTx && result.MultiportBasePort != via.UdpAddr.Port() {
// The other side sent us a handshake from a different port, make sure
// we send responses back to the BasePort
via.UdpAddr = netip.AddrPortFrom(via.UdpAddr.Addr(), result.MultiportBasePort)
}
hostinfo.multiportTx = hm.multiPort.Tx && result.MultiportRx
hostinfo.multiportRx = hm.multiPort.Rx && result.MultiportTx
// Handshake complete; build the ConnectionState now that we have keys and a verified peer cert.
hostinfo.ConnectionState = newConnectionStateFromResult(result)
@@ -1010,8 +961,6 @@ func (hm *HandshakeManager) continueHandshake(via ViaSender, hh *HandshakeHostIn
"handshake", m{"stage": uint64(machine.MessageIndex()), "style": header.SubTypeName(header.Handshake, machine.Subtype())},
"durationNs", duration,
"sentCachedPackets", len(hh.packetStore),
"multiportTx", hostinfo.multiportTx,
"multiportRx", hostinfo.multiportRx,
)
hostinfo.vpnAddrs = vpnAddrs
@@ -1153,10 +1102,6 @@ func (hm *HandshakeManager) handleCheckAndCompleteError(err error, existing, hos
switch err {
case ErrAlreadySeen:
if hostinfo.multiportRx {
// The other host is sending to us with multiport, so only grab the IP
via.UdpAddr = netip.AddrPortFrom(via.UdpAddr.Addr(), hostinfo.GetRemote().Port())
}
if existing.SetRemoteIfPreferred(f.hostMap, via) {
f.SendMessageToVpnAddr(header.Test, header.TestRequest, hostinfo.vpnAddrs[0], []byte(""), make([]byte, 12, 12), make([]byte, mtu))
}
+27 -10
View File
@@ -254,12 +254,6 @@ type HostInfo struct {
networks *bart.Table[NetworkType]
relayState RelayState
// If true, we should send to this remote using multiport
multiportTx bool
// If true, we will receive from this remote using multiport
multiportRx bool
// HandshakePacket records the packets used to create this hostinfo
// We need these to avoid replayed handshake packets creating new hostinfos which causes churn
HandshakePacket map[uint8][]byte
@@ -874,10 +868,28 @@ func (i *HostInfo) logger(l *slog.Logger) *slog.Logger {
// Utility functions
func localAddrs(l *slog.Logger, allowList *LocalAllowList) []netip.Addr {
func localAddrs(l *slog.Logger, allowList *LocalAllowList) ([]netip.Addr, error) {
return collectLocalAddrs(l, allowList, localInterfaces, localInterfaceAddrs)
}
// collectLocalAddrs takes its enumerators as arguments so tests can drive the filtering and the
// failure branches without depending on the addresses of whatever host they run on. It reports
// failures to the caller rather than logging them, because it runs on every lighthouse update and
// only the caller can tell a new failure from a repeat of the same one.
func collectLocalAddrs(
l *slog.Logger,
allowList *LocalAllowList,
interfaces func() ([]net.Interface, error),
interfaceAddrs func(*net.Interface) ([]net.Addr, error),
) ([]netip.Addr, error) {
//FIXME: This function is pretty garbage
var finalAddrs []netip.Addr
ifaces, _ := net.Interfaces()
var errs []error
ifaces, err := interfaces()
if err != nil {
return nil, fmt.Errorf("failed to enumerate local interfaces: %w", err)
}
for _, i := range ifaces {
allow := allowList.AllowName(i.Name)
if l.Enabled(context.Background(), logging.LevelTrace) {
@@ -890,7 +902,12 @@ func localAddrs(l *slog.Logger, allowList *LocalAllowList) []netip.Addr {
if !allow {
continue
}
addrs, _ := i.Addrs()
addrs, err := interfaceAddrs(&i)
if err != nil {
errs = append(errs, fmt.Errorf("failed to get addresses for %s: %w", i.Name, err))
continue
}
for _, rawAddr := range addrs {
var addr netip.Addr
switch v := rawAddr.(type) {
@@ -925,5 +942,5 @@ func localAddrs(l *slog.Logger, allowList *LocalAllowList) []netip.Addr {
}
}
}
return finalAddrs
return finalAddrs, errors.Join(errs...)
}
+82
View File
@@ -1,6 +1,8 @@
package nebula
import (
"errors"
"net"
"net/netip"
"slices"
"testing"
@@ -401,3 +403,83 @@ func TestHostMap_RelayState(t *testing.T) {
assert.Equal(t, []netip.Addr{}, h1.relayState.relays)
}
func TestCollectLocalAddrs(t *testing.T) {
ifaces := []net.Interface{
{Index: 1, Name: "lo"},
{Index: 2, Name: "eth0"},
{Index: 3, Name: "docker0"},
}
addrs := map[string][]net.Addr{
"lo": {
&net.IPNet{IP: net.ParseIP("127.0.0.1"), Mask: net.CIDRMask(8, 32)},
&net.IPNet{IP: net.ParseIP("::1"), Mask: net.CIDRMask(128, 128)},
},
"eth0": {
&net.IPNet{IP: net.ParseIP("10.0.0.5"), Mask: net.CIDRMask(24, 32)},
&net.IPNet{IP: net.ParseIP("fe80::1"), Mask: net.CIDRMask(64, 128)},
&net.IPAddr{IP: net.ParseIP("fd00::5")},
},
"docker0": {
&net.IPNet{IP: net.ParseIP("172.17.0.1"), Mask: net.CIDRMask(16, 32)},
},
}
enumerate := func() ([]net.Interface, error) { return ifaces, nil }
addrsFor := func(i *net.Interface) ([]net.Addr, error) { return addrs[i.Name], nil }
// Loopback and link local are dropped, everything else on every interface is kept.
out, err := collectLocalAddrs(test.NewLogger(), nil, enumerate, addrsFor)
require.NoError(t, err)
assert.Equal(t, []netip.Addr{
netip.MustParseAddr("10.0.0.5"),
netip.MustParseAddr("fd00::5"),
netip.MustParseAddr("172.17.0.1"),
}, out)
// An interface the allow list rejects by name is never asked for its addresses.
c := config.NewC(test.NewLogger())
c.Settings["allowlist"] = map[string]any{
"interfaces": map[string]any{`docker.*`: false},
}
al, err := NewLocalAllowListFromConfig(c, "allowlist")
require.NoError(t, err)
asked := make(map[string]struct{})
countingAddrsFor := func(i *net.Interface) ([]net.Addr, error) {
asked[i.Name] = struct{}{}
return addrs[i.Name], nil
}
out, err = collectLocalAddrs(test.NewLogger(), al, enumerate, countingAddrsFor)
require.NoError(t, err)
assert.Equal(t, []netip.Addr{
netip.MustParseAddr("10.0.0.5"),
netip.MustParseAddr("fd00::5"),
}, out)
assert.NotContains(t, asked, "docker0")
// A failure to enumerate interfaces at all is reported rather than silently advertising nothing.
out, err = collectLocalAddrs(
test.NewLogger(),
nil,
func() ([]net.Interface, error) { return nil, errors.New("netlinkrib: permission denied") },
addrsFor,
)
assert.Nil(t, out)
require.EqualError(t, err, "failed to enumerate local interfaces: netlinkrib: permission denied")
// One interface failing is reported and skipped, the rest are still collected.
out, err = collectLocalAddrs(
test.NewLogger(),
nil,
enumerate,
func(i *net.Interface) ([]net.Addr, error) {
if i.Name == "eth0" {
return nil, errors.New("nope")
}
return addrs[i.Name], nil
},
)
assert.Equal(t, []netip.Addr{netip.MustParseAddr("172.17.0.1")}, out)
require.EqualError(t, err, "failed to get addresses for eth0: nope")
}
+8 -38
View File
@@ -10,7 +10,6 @@ import (
"github.com/slackhq/nebula/iputil"
"github.com/slackhq/nebula/noiseutil"
"github.com/slackhq/nebula/routing"
"github.com/slackhq/nebula/udp"
)
func (f *Interface) consumeInsidePacket(packet []byte, fwPacket *firewall.Packet, nb, out []byte, q int, localCache firewall.ConntrackCache) {
@@ -74,7 +73,7 @@ func (f *Interface) consumeInsidePacket(packet []byte, fwPacket *firewall.Packet
dropReason := f.firewall.Drop(*fwPacket, false, hostinfo, f.pki.GetCAPool(), localCache)
if dropReason == nil {
f.sendNoMetrics(header.Message, 0, hostinfo.ConnectionState, hostinfo, netip.AddrPort{}, packet, nb, out, q, fwPacket)
f.sendNoMetrics(header.Message, 0, hostinfo.ConnectionState, hostinfo, netip.AddrPort{}, packet, nb, out, q)
} else {
f.rejectInside(packet, out, q)
@@ -123,7 +122,7 @@ func (f *Interface) rejectOutside(packet []byte, ci *ConnectionState, hostinfo *
return
}
f.sendNoMetrics(header.Message, 0, ci, hostinfo, netip.AddrPort{}, out, nb, packet, q, nil)
f.sendNoMetrics(header.Message, 0, ci, hostinfo, netip.AddrPort{}, out, nb, packet, q)
}
// Handshake will attempt to initiate a tunnel with the provided vpn address. This is a no-op if the tunnel is already established or being established
@@ -236,7 +235,7 @@ func (f *Interface) sendMessageNow(t header.MessageType, st header.MessageSubTyp
return
}
f.sendNoMetrics(header.Message, st, hostinfo.ConnectionState, hostinfo, netip.AddrPort{}, p, nb, out, 0, nil)
f.sendNoMetrics(header.Message, st, hostinfo.ConnectionState, hostinfo, netip.AddrPort{}, p, nb, out, 0)
}
// SendMessageToVpnAddr handles real addr:port lookup and sends to the current best known address for vpnAddr.
@@ -268,12 +267,12 @@ func (f *Interface) SendMessageToHostInfo(t header.MessageType, st header.Messag
func (f *Interface) send(t header.MessageType, st header.MessageSubType, ci *ConnectionState, hostinfo *HostInfo, p, nb, out []byte) {
f.messageMetrics.Tx(t, st, 1)
f.sendNoMetrics(t, st, ci, hostinfo, netip.AddrPort{}, p, nb, out, 0, nil)
f.sendNoMetrics(t, st, ci, hostinfo, netip.AddrPort{}, p, nb, out, 0)
}
func (f *Interface) sendTo(t header.MessageType, st header.MessageSubType, ci *ConnectionState, hostinfo *HostInfo, remote netip.AddrPort, p, nb, out []byte) {
f.messageMetrics.Tx(t, st, 1)
f.sendNoMetrics(t, st, ci, hostinfo, remote, p, nb, out, 0, nil)
f.sendNoMetrics(t, st, ci, hostinfo, remote, p, nb, out, 0)
}
// SendVia sends a payload through a Relay tunnel. No authentication or encryption is done
@@ -341,27 +340,10 @@ func (f *Interface) SendVia(via *HostInfo,
f.connectionManager.RelayUsed(relay.LocalIndex)
}
func (f *Interface) sendNoMetrics(t header.MessageType, st header.MessageSubType, ci *ConnectionState, hostinfo *HostInfo, remote netip.AddrPort, p, nb, out []byte, q int, udpPortGetter udp.SendPortGetter) {
func (f *Interface) sendNoMetrics(t header.MessageType, st header.MessageSubType, ci *ConnectionState, hostinfo *HostInfo, remote netip.AddrPort, p, nb, out []byte, q int) {
if ci.eKey == nil {
return
}
multiport := f.multiPort.Tx && hostinfo.multiportTx
rawOut := out
if multiport {
if len(out) < udp.RawOverhead {
// NOTE: This is because some spots in the code send us `out[:0]`, so
// we need to expand the slice back out to get our 8 bytes back.
out = out[:udp.RawOverhead]
}
// Preserve bytes needed for the raw socket
out = out[udp.RawOverhead:]
if udpPortGetter == nil {
udpPortGetter = udp.RandomSendPort
}
}
useRelay := !remote.IsValid() && !hostinfo.GetRemote().IsValid()
fullOut := out
@@ -414,13 +396,7 @@ func (f *Interface) sendNoMetrics(t header.MessageType, st header.MessageSubType
}
if remote.IsValid() {
if multiport {
rawOut = rawOut[:len(out)+udp.RawOverhead]
port := udpPortGetter.UDPSendPort(f.multiPort.TxPorts)
err = f.udpRaw.WriteTo(rawOut, port, remote)
} else {
err = f.writers[q].WriteTo(out, remote)
}
err = f.writers[q].WriteTo(out, remote)
if err != nil {
hostinfo.logger(f.l).Error("Failed to write outgoing packet",
"error", err,
@@ -428,13 +404,7 @@ func (f *Interface) sendNoMetrics(t header.MessageType, st header.MessageSubType
)
}
} else if hr := hostinfo.GetRemote(); hr.IsValid() {
if multiport {
rawOut = rawOut[:len(out)+udp.RawOverhead]
port := udpPortGetter.UDPSendPort(f.multiPort.TxPorts)
err = f.udpRaw.WriteTo(rawOut, port, hr)
} else {
err = f.writers[q].WriteTo(out, hr)
}
err = f.writers[q].WriteTo(out, hr)
if err != nil {
hostinfo.logger(f.l).Error("Failed to write outgoing packet",
"error", err,
-23
View File
@@ -99,9 +99,6 @@ type Interface struct {
// triggerShutdown is a function that will be run exactly once, when onFatal swaps something non-nil into fatalErr
triggerShutdown func()
udpRaw *udp.RawConn
multiPort config.MultiPortConfig
metricHandshakes metrics.Histogram
messageMetrics *MessageMetrics
cachedPacketMetrics *cachedPacketMetrics
@@ -109,15 +106,6 @@ type Interface struct {
l *slog.Logger
}
type MultiPortConfig struct {
Tx bool
Rx bool
TxBasePort uint16
TxPorts int
TxHandshake bool
TxHandshakeDelay int64
}
type EncWriter interface {
SendVia(via *HostInfo,
relay *Relay,
@@ -261,8 +249,6 @@ func (f *Interface) activate() error {
metrics.GetOrRegisterGauge("routines", nil).Update(int64(f.routines))
metrics.GetOrRegisterGauge("multiport.tx_ports", nil).Update(int64(f.multiPort.TxPorts))
// Prepare n tun queues
var reader io.ReadWriteCloser = f.inside
for i := 0; i < f.routines; i++ {
@@ -519,8 +505,6 @@ func (f *Interface) emitStats(ctx context.Context, i time.Duration) {
udpStats := udp.NewUDPStatsEmitter(f.writers)
var rawStats func()
certExpirationGauge := metrics.GetOrRegisterGauge("certificate.ttl_seconds", nil)
certInitiatingVersion := metrics.GetOrRegisterGauge("certificate.initiating_version", nil)
certMaxVersion := metrics.GetOrRegisterGauge("certificate.max_version", nil)
@@ -535,13 +519,6 @@ func (f *Interface) emitStats(ctx context.Context, i time.Duration) {
certExpirationGauge.Update(int64(defaultCrt.NotAfter().Sub(time.Now()) / time.Second))
certInitiatingVersion.Update(int64(defaultCrt.Version()))
if f.udpRaw != nil {
if rawStats == nil {
rawStats = udp.NewRawStatsEmitter(f.udpRaw)
}
rawStats()
}
// Report the max certificate version we are capable of using
if certState.v2Cert != nil {
certMaxVersion.Update(int64(certState.v2Cert.Version()))
+27 -1
View File
@@ -40,6 +40,10 @@ type LightHouse struct {
// addresses rather than whatever this machine's NICs happen to be. Set it before Start.
localAddrsFn func(*LocalAllowList) []netip.Addr
// lastLocalAddrsErr is the previous localAddrsFn failure. Enumeration runs on every update, so an
// unchanged failure is demoted to Debug rather than warning every lighthouse.interval forever.
lastLocalAddrsErr atomic.Pointer[string]
// Local cache of answers from light houses
// map of vpn addr to answers
addrMap map[netip.Addr]*RemoteList
@@ -112,7 +116,9 @@ func NewLightHouseFromConfig(ctx context.Context, l *slog.Logger, c *config.C, c
l: l,
}
h.localAddrsFn = func(al *LocalAllowList) []netip.Addr {
return localAddrs(h.l, al)
addrs, err := localAddrs(h.l, al)
h.logLocalAddrsErr(err)
return addrs
}
lighthouses := make([]netip.Addr, 0)
@@ -913,6 +919,26 @@ func (lh *LightHouse) TriggerUpdate() {
}
}
// logLocalAddrsErr reports a localAddrs failure at Warn the first time it is seen and at Debug while
// it persists unchanged, so a permanent failure does not warn on every update forever.
func (lh *LightHouse) logLocalAddrsErr(err error) {
if err == nil {
lh.lastLocalAddrsErr.Store(nil)
return
}
msg := err.Error()
prev := lh.lastLocalAddrsErr.Swap(&msg)
if prev != nil && *prev == msg {
if lh.l.Enabled(context.Background(), slog.LevelDebug) {
lh.l.Debug("Failed to collect local addresses to advertise", "error", err)
}
return
}
lh.l.Warn("Failed to collect local addresses to advertise", "error", err)
}
func (lh *LightHouse) SendUpdate() {
var v4 []*V4AddrPort
var v6 []*V6AddrPort
+31
View File
@@ -1,7 +1,9 @@
package nebula
import (
"bytes"
"encoding/binary"
"errors"
"fmt"
"net/netip"
"testing"
@@ -738,3 +740,32 @@ func TestLighthouse_DeletesWork(t *testing.T) {
out = lh.Query(testHost)
assert.Nil(t, out)
}
func TestLightHouse_logLocalAddrsErr(t *testing.T) {
out := &bytes.Buffer{}
lh := &LightHouse{l: test.NewLoggerWithOutput(out)}
// The first sighting of a failure warns.
lh.logLocalAddrsErr(errors.New("permission denied"))
assert.Contains(t, out.String(), "level=WARN")
assert.Contains(t, out.String(), "permission denied")
// Repeating unchanged does not warn again, which is what keeps a permanent failure from warning
// on every lighthouse.interval for the life of the process.
out.Reset()
lh.logLocalAddrsErr(errors.New("permission denied"))
assert.NotContains(t, out.String(), "level=WARN")
// A different failure is a new event and warns.
out.Reset()
lh.logLocalAddrsErr(errors.New("something else"))
assert.Contains(t, out.String(), "level=WARN")
assert.Contains(t, out.String(), "something else")
// Recovering resets, so the same failure returning later warns again.
out.Reset()
lh.logLocalAddrsErr(nil)
assert.Empty(t, out.String())
lh.logLocalAddrsErr(errors.New("something else"))
assert.Contains(t, out.String(), "level=WARN")
}
+13
View File
@@ -0,0 +1,13 @@
//go:build !android
package nebula
import "net"
func localInterfaces() ([]net.Interface, error) {
return net.Interfaces()
}
func localInterfaceAddrs(i *net.Interface) ([]net.Addr, error) {
return i.Addrs()
}
+32
View File
@@ -0,0 +1,32 @@
//go:build android
package nebula
import (
"net"
"github.com/wlynxg/anet"
)
// anet relies on //go:linkname and so needs -ldflags=-checklinkname=0 on Go 1.23+. Nebula ships no
// Android binaries of its own, so that burden falls on consumers linking Android artifacts.
func init() {
// anet only takes its bind-free path when it believes it is on API 30+, and detecting the running
// device's level requires cgo. Pin it so a CGO_ENABLED=0 build cannot quietly fall back to the
// denied path. The bind-free path is correct on older releases too, just unnecessary there.
anet.SetAndroidVersion(11)
}
// The app sandbox denies bind() on netlink_route_socket, so the stdlib's RTM_GETLINK enumeration
// fails with EACCES and we advertise no underlay addresses at all. anet reads RTM_GETADDR from an
// unbound socket instead, so this must not be collapsed back into net.Interfaces.
func localInterfaces() ([]net.Interface, error) {
return anet.Interfaces()
}
// net.Interface.Addrs goes back through the denied netlink path, so addresses have to come from anet
// as well. anet cannot report HardwareAddr, which localAddrs does not read.
func localInterfaceAddrs(i *net.Interface) ([]net.Addr, error) {
return anet.InterfaceAddrsByInterface(i)
}
-33
View File
@@ -244,39 +244,6 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
ifce.writers = udpConns
lightHouse.ifce = ifce
loadMultiPortConfig := func(c *config.C) {
ifce.multiPort.Rx = c.GetBool("tun.multiport.rx_enabled", false)
tx := c.GetBool("tun.multiport.tx_enabled", false)
if tx && ifce.udpRaw == nil {
ifce.udpRaw, err = udp.NewRawConn(l, c.GetString("listen.host", "0.0.0.0"), port, uint16(port))
if err != nil {
l.Error("Failed to get raw socket for tun.multiport.tx_enabled", "error", err)
ifce.udpRaw = nil
tx = false
}
}
if tx {
ifce.multiPort.TxBasePort = uint16(port)
ifce.multiPort.TxPorts = c.GetInt("tun.multiport.tx_ports", 100)
ifce.multiPort.TxHandshake = c.GetBool("tun.multiport.tx_handshake", false)
ifce.multiPort.TxHandshakeDelay = int64(c.GetInt("tun.multiport.tx_handshake_delay", 2))
ifce.udpRaw.ReloadConfig(c)
}
ifce.multiPort.Tx = tx
// TODO: if we upstream this, make this cleaner
handshakeManager.udpRaw = ifce.udpRaw
handshakeManager.multiPort = ifce.multiPort
l.Info("Multiport configured", "multiPort", ifce.multiPort)
}
loadMultiPortConfig(c)
c.RegisterReloadCallback(loadMultiPortConfig)
ifce.RegisterConfigChangeCallbacks(c)
ifce.reloadDisconnectInvalid(c)
ifce.reloadSendRecvError(c)
-9
View File
@@ -264,15 +264,6 @@ func (f *Interface) sendCloseTunnel(h *HostInfo) {
func (f *Interface) handleHostRoaming(hostinfo *HostInfo, via ViaSender) {
curRemote := hostinfo.GetRemote()
if !via.IsRelayed && curRemote != via.UdpAddr {
if hostinfo.multiportRx {
// If the remote is sending with multiport, we aren't roaming unless
// the IP has changed
if curRemote.Addr().Compare(via.UdpAddr.Addr()) == 0 {
return
}
// Keep the port from the original hostinfo, because the remote is transmitting from multiport ports
via.UdpAddr = netip.AddrPortFrom(via.UdpAddr.Addr(), curRemote.Port())
}
if !f.lightHouse.GetRemoteAllowList().AllowAll(hostinfo.vpnAddrs, via.UdpAddr.Addr()) {
if f.l.Enabled(context.Background(), slog.LevelDebug) {
hostinfo.logger(f.l).Debug("lighthouse.remote_allow_list denied roaming", "newAddr", via.UdpAddr)
-16
View File
@@ -1,16 +0,0 @@
package udp
import mathrand "math/rand"
type SendPortGetter interface {
// UDPSendPort returns the port to use
UDPSendPort(maxPort int) uint16
}
type randomSendPort struct{}
func (randomSendPort) UDPSendPort(maxPort int) uint16 {
return uint16(mathrand.Intn(maxPort))
}
var RandomSendPort = randomSendPort{}
-191
View File
@@ -1,191 +0,0 @@
//go:build !android && !e2e_testing
// +build !android,!e2e_testing
package udp
import (
"encoding/binary"
"fmt"
"log/slog"
"net"
"net/netip"
"syscall"
"unsafe"
"github.com/rcrowley/go-metrics"
"github.com/slackhq/nebula/config"
"golang.org/x/net/ipv4"
"golang.org/x/sys/unix"
)
// RawOverhead is the number of bytes that need to be reserved at the start of
// the raw bytes passed to (*RawConn).WriteTo. This is used by WriteTo to prefix
// the IP and UDP headers.
const RawOverhead = 28
type RawConn struct {
sysFd int
basePort uint16
l *slog.Logger
}
func NewRawConn(l *slog.Logger, ip string, port int, basePort uint16) (*RawConn, error) {
syscall.ForkLock.RLock()
// With IPPROTO_UDP, the linux kernel tries to deliver every UDP packet
// received in the system to our socket. This constantly overflows our
// buffer and marks our socket as having dropped packets. This makes the
// stats on the socket useless.
//
// In contrast, IPPROTO_RAW is not delivered any packets and thus our read
// buffer will not fill up and mark as having dropped packets. The only
// difference is that we have to assemble the IP header as well, but this
// is fairly easy since Linux does the checksum for us.
//
// TODO: How to get this working with Inet6 correctly? I was having issues
// with the source address when testing before, probably need to `bind(2)`?
fd, err := unix.Socket(unix.AF_INET, unix.SOCK_RAW, unix.IPPROTO_RAW)
if err == nil {
unix.CloseOnExec(fd)
}
syscall.ForkLock.RUnlock()
if err != nil {
return nil, err
}
// We only want to send, not recv. This will hopefully help the kernel avoid
// wasting time on us
if err = unix.SetsockoptInt(fd, unix.SOL_SOCKET, unix.SO_RCVBUF, 0); err != nil {
return nil, fmt.Errorf("unable to set SO_RCVBUF: %s", err)
}
var lip [16]byte
copy(lip[:], net.ParseIP(ip))
// TODO do we need to `bind(2)` so that we send from the correct address/interface?
if err = unix.Bind(fd, &unix.SockaddrInet6{Addr: lip, Port: port}); err != nil {
return nil, fmt.Errorf("unable to bind to socket: %s", err)
}
return &RawConn{
sysFd: fd,
basePort: basePort,
l: l,
}, nil
}
// WriteTo must be called with raw leaving the first `udp.RawOverhead` bytes empty,
// for the IP/UDP headers.
func (u *RawConn) WriteTo(raw []byte, fromPort uint16, ip netip.AddrPort) error {
var rsa unix.RawSockaddrInet4
rsa.Family = unix.AF_INET
rsa.Addr = ip.Addr().As4()
totalLen := len(raw)
udpLen := totalLen - ipv4.HeaderLen
// IP header
raw[0] = byte(ipv4.Version<<4 | (ipv4.HeaderLen >> 2 & 0x0f))
raw[1] = 0 // tos
binary.BigEndian.PutUint16(raw[2:4], uint16(totalLen))
binary.BigEndian.PutUint16(raw[4:6], 0) // id (linux does it for us)
binary.BigEndian.PutUint16(raw[6:8], 0) // frag options
raw[8] = byte(64) // ttl
raw[9] = byte(17) // protocol
binary.BigEndian.PutUint16(raw[10:12], 0) // checksum (linux does it for us)
binary.BigEndian.PutUint32(raw[12:16], 0) // src (linux does it for us)
copy(raw[16:20], rsa.Addr[:]) // dst
// UDP header
fromPort = u.basePort + fromPort
binary.BigEndian.PutUint16(raw[20:22], uint16(fromPort)) // src port
binary.BigEndian.PutUint16(raw[22:24], uint16(ip.Port())) // dst port
binary.BigEndian.PutUint16(raw[24:26], uint16(udpLen)) // UDP length
binary.BigEndian.PutUint16(raw[26:28], 0) // checksum (optional)
for {
_, _, err := unix.Syscall6(
unix.SYS_SENDTO,
uintptr(u.sysFd),
uintptr(unsafe.Pointer(&raw[0])),
uintptr(len(raw)),
uintptr(0),
uintptr(unsafe.Pointer(&rsa)),
uintptr(unix.SizeofSockaddrInet4),
)
if err != 0 {
return &net.OpError{Op: "sendto", Err: err}
}
//TODO: handle incomplete writes
return nil
}
}
func (u *RawConn) ReloadConfig(c *config.C) {
b := c.GetInt("listen.write_buffer", 0)
if b <= 0 {
return
}
if err := u.SetSendBuffer(b); err != nil {
u.l.Error("Failed to set listen.write_buffer", "error", err)
return
}
s, err := u.GetSendBuffer()
if err != nil {
u.l.Warn("Failed to get listen.write_buffer", "error", err)
return
}
u.l.Info("listen.write_buffer was set", "size", s)
}
func (u *RawConn) SetSendBuffer(n int) error {
return unix.SetsockoptInt(u.sysFd, unix.SOL_SOCKET, unix.SO_SNDBUFFORCE, n)
}
func (u *RawConn) GetSendBuffer() (int, error) {
return unix.GetsockoptInt(u.sysFd, unix.SOL_SOCKET, unix.SO_SNDBUF)
}
func (u *RawConn) getMemInfo(meminfo *[unix.SK_MEMINFO_VARS]uint32) error {
var vallen uint32 = 4 * unix.SK_MEMINFO_VARS
_, _, err := unix.Syscall6(unix.SYS_GETSOCKOPT, uintptr(u.sysFd), uintptr(unix.SOL_SOCKET), uintptr(unix.SO_MEMINFO), uintptr(unsafe.Pointer(meminfo)), uintptr(unsafe.Pointer(&vallen)), 0)
if err != 0 {
return err
}
return nil
}
func NewRawStatsEmitter(rawConn *RawConn) func() {
// Check if our kernel supports SO_MEMINFO before registering the gauges
var gauges [unix.SK_MEMINFO_VARS]metrics.Gauge
var meminfo [unix.SK_MEMINFO_VARS]uint32
if err := rawConn.getMemInfo(&meminfo); err == nil {
gauges = [unix.SK_MEMINFO_VARS]metrics.Gauge{
metrics.GetOrRegisterGauge("raw.rmem_alloc", nil),
metrics.GetOrRegisterGauge("raw.rcvbuf", nil),
metrics.GetOrRegisterGauge("raw.wmem_alloc", nil),
metrics.GetOrRegisterGauge("raw.sndbuf", nil),
metrics.GetOrRegisterGauge("raw.fwd_alloc", nil),
metrics.GetOrRegisterGauge("raw.wmem_queued", nil),
metrics.GetOrRegisterGauge("raw.optmem", nil),
metrics.GetOrRegisterGauge("raw.backlog", nil),
metrics.GetOrRegisterGauge("raw.drops", nil),
}
} else {
// return no-op because we don't support SO_MEMINFO
return func() {}
}
return func() {
if err := rawConn.getMemInfo(&meminfo); err == nil {
for j := 0; j < unix.SK_MEMINFO_VARS; j++ {
gauges[j].Update(int64(meminfo[j]))
}
}
}
}
-29
View File
@@ -1,29 +0,0 @@
//go:build !linux || android || e2e_testing
// +build !linux android e2e_testing
package udp
import (
"fmt"
"log/slog"
"net/netip"
"runtime"
"github.com/slackhq/nebula/config"
)
const RawOverhead = 0
type RawConn struct{}
func NewRawConn(l *slog.Logger, ip string, port int, basePort uint16) (*RawConn, error) {
return nil, fmt.Errorf("multiport tx is not supported on %s", runtime.GOOS)
}
func (u *RawConn) WriteTo(raw []byte, fromPort uint16, addr netip.AddrPort) error {
return fmt.Errorf("multiport tx is not supported on %s", runtime.GOOS)
}
func (u *RawConn) ReloadConfig(c *config.C) {}
func NewRawStatsEmitter(rawConn *RawConn) func() { return func() {} }